+
+ * config/tc-hppa.c (hppa_fix_adjustable): Reject reductions for
+ symbols in DLT relative relocs.
+ (tc_gen_reloc): Zero out the addend field for DLT relative relocs.
+
+
+ * config/tc-ppc.c (ppc_tc): If not OBJ_COFF, force TOC entry to
+ align to a four byte boundary.
+
+
+ * config/tc-alpha.c (load_expression): Handle 32-bit addends.
+ (gpdisp_hi16_howto): Now points to const.
+ (load_insn_table, alpha_ip): Fix uses of const.
+
+ * doc/internals.texi: Updates to COFF description. Added "@end
+ defmac" as needed, and some extra heading and "@bye" so it'll
+ format as a separate document.
+
+
+ * config/obj-elf.h (S_GET_ALIGN, S_SET_ALIGN): Define.
+ * config/obj-elf.c (obj_elf_common): Set alignment of common
+ symbol.
+ * config/tc-sparc.c (s_common): If OBJ_ELF, set alignment of
+ common symbol.
+
+
+ * config/tc-hppa.c (is_complex): New macro.
+ (cons_fix_new_hppa): "Handle" complex expressions.
+
+
+ * config/obj-coff.c (write_object_file): Set s_align field from
+ section_alignment array.
+
+
+ * write.c (write_object_file): Use bfd_com_section_ptr.
+ * as.h (absolute_section, undefined_section): Use new BFD macros
+ bfd_abs_section_ptr and bfd_und_section_ptr.
+
+
+ * config/tc-mips.c (macro): For M_LI_SS, decide how to handle it
+ based on contents of imm_expr and offset_expr, rather than
+ mips_pic. For M_LI_DD, decide how to handle it based on segment
+ name of offset_expr, rather than mips_pic.
+ (mips_ip): If g_switch_value < 4, use immediate values for 'l'.
+ If g_switch_value < 8, use .rdata rather than .lit for 'L'.
+
+
+ * as.c (show_usage): Break long string into shorter ones.
+ (parse_args): Add -v, prints version id and continues.
+ * config/tc-mips.c (md_show_usage): Break long string.
+
+
+ * config/tc-i386.c (md_parse_option): Handle "-V" and "-Q" if
+ OBJ_ELF is defined.
+
+
+ * as.c (main) [HOST_SPECIAL_INIT]: New hook, for host-specific
+ initialization.
+
+
+ * config/tc-hppa.h (obj_frob_symbol): Define for OBJ_ELF.
+ More gas/bfd lossage exposed by the new linker code.
+
+
+ * subsegs.c (subseg_change): Clear seginfo after allocating it.
+ (subseg_get): Pass actual size of seginfo to memset.
+
+ * subsegs.c (abs_seg_info, und_seg_info): Define if BFD_ASSEMBLER.
+ (subseg_change): Store seg_info for bfd_abs_section_ptr in
+ abs_seg_info, and store seg_info for bfd_und_section_ptr in
+ und_seg_info.
+ (subseg_get): Likewise. Also, don't set output_section if it is
+ already set.
+ (seg_info): Define as function.
+ * subsegs.h (seg_info): Declare as function rather than defining
+ as macro.
+ * write.c (relax_and_size_seg): Call seg_info rather than
+ bfd_get_section_userdata.
+
+
+ * configure.in (ppc-*-elf*): New target, like -sysv4*.
+
+ * expr.c (operand): If "0f" is followed by '\0', don't do eol
+ checks.
+
* ecoff.c (ecoff_build_aux): Call swap_tir_out and swap_rndx_out
(DBG_S_C_REAL8_G, DBG_S_C_COMPLX8_G): G_float versions of
REAL8 and COMPLX8; not used yet, because gcc outputs the same
.stabs for `double' regardless of whether `-mg' is used.
- config/obj-vms.c (VMS_typedef_parse) [case 'r']: add entries
+ * config/obj-vms.c (VMS_typedef_parse) [case 'r']: add entries
for gcc2's predefined types "complex float", "complex double",
and "complex long double" (identical to complex double).
for `long long' support. Their values come from the existing
DSC$K_DTYPE_QU and DSC$K_DTYPE_Q macros in <descrip.h>. The
VMS debugger now recognizes `long long' variables correctly.
- config/obj-vms.c (VMS_typedef_parse) [case 'r']: add entries
+ * config/obj-vms.c (VMS_typedef_parse) [case 'r']: add entries
for gcc2's predefined types "long double" (same as double, as
per gcc's current state), "long long int", "long long unsigned
int", and final `otherwise' case (to avoid uninitialized type